Merge pull request #354 from PyThaiNLP/verb-commands

Verb-like commands for "thainlp" command-line program

Author: bact

docs/notes/command_line.rst

@@ -1,52 +1,75 @@
 Command Line
 ============
 
-You can use PyThaiNLP from Command Line.
+You can use some thainlp functions directly from command line.
 
-Run Command Line and type the commands:::
+**Tokenization**::
 
-    pythainlp
+    thainlp tokenize <word|syllable|subword|sent> [-w] [-nw] [-a newmm|attacut|longest] [-s SEPARATOR] TEXT
 
-**Word tokenization**::
+*Example*::
 
-    pythainlp tokenization word --text TEXT
+    $ thainlp tokenize word สภาพการจ้างและสภาพการทำงาน
+    สภาพการจ้าง|และ|สภาพ|การทำงาน|
 
-*Example*::
+    $ thainlp tokenize syllable สภาพการจ้างและสภาพการทำงาน
+    สภาพ~การ~จ้าง~และ~สภาพ~การ~ทำ~งาน~
 
-    $ pythainlp tokenization word --text "ผมร<0e31>กประเทศไทย? สามารถ" --engine newmm
-    ผม|รัก|ประเทศไทย|?| |สามารถ
+    $ thainlp tokenize subword สภาพการจ้างและสภาพการทำงาน
+    ส/ภา/พ/กา/ร/จ้า/ง/และ/ส/ภา/พ/กา/ร/ทำ/งา/น/
 
+    $ thainlp tokenize word -a longest "แรงงานกะดึก: ฟันเฟืองที่ยังหมุนในคำ่คืนมีเคอร์ฟิว"
+    แรงงาน|กะ|ดึก|:| |ฟันเฟือง|ที่|ยัง|หมุน|ใน|คำ่|คืน|มี|เคอร์ฟิว|
 
-**Syllable tokenization**::
+    $ thainlp tokenize word -nw -s # "5 เหตุผล 'ไม่ควร' ต่อพ.ร.ก.ฉุกเฉิน"
+    5#เหตุผล#'#ไม่#ควร#'#ต่อ#พ.ร.ก.#ฉุกเฉิน#
 
-    pythainlp tokenization syllable --text TEXT
+**Part-Of-Speech tagging**::
+
+    pythainlp tagg pos [-s SEPARATOR] TEXT
 
 *Example*::
 
-    $ pythainlp tokenization syllable --text "ผมร<0e31>กประเทศไทย? สามารถ"
-    ผม~รัก~ประ~เทศ~ไทย~? ~สา~มารถ
+    $ thainlp tag pos -s . ผม.ไม่.กิน.เผ็ด
 
-**Part-Of-Speech tagging**::
+**Soundex**::
 
-    pythainlp tagging pos --text TEXT
+    thainlp soundex [-a udom83|lk82|metasound] TEXT
 
 *Example*::
 
-    $ pythainlp tagging pos --text "ผม|ไม่|กิน|เผ็ด"
+    $ thainlp soundex วรรณ
+    ว330000
 
-**Soundex**::
+    $ thainlp soundex -a lk82 วัน
+    ว4000
+
+    $ thainlp soundex -a lk82 วรรณ
+    ว4000
 
-    pythainlp soundex --text TEXT
+**Corpus management**::
+
+    thainlp data <catalog|info|get|rm|path>
 
 *Example*::
 
-    $ pythainlp soundex --text "บ<0e39>รณการ" --engine lk82
-    บE419
+    $ thainlp data path
+    /Users/user1/pythainlp-data
+
+    $ thainlp data catalog
+    Dataset/corpus available for download:
+    - crfcut 0.1
+    - thai-g2p 0.1  (Local: 0.1)
+    - thai2fit_wv 0.1
+    - thainer-1-3 1.3
 
-**Mange corpus**::
+    $ thainlp data get thai2fit_wv
+    Corpus: thai2fit_wv
+    - Downloading: thai2fit_wv 0.1
+    36%|█████████████████▉                                |
 
-    pythainlp corpus
+    $ thainlp data --help
 
 **Help**::
 
-    pythainlp --help
+    thainlp --help

docs/notes/installation.rst

@@ -50,5 +50,8 @@ Runtime Configurations
    This environment variable specifies the location where the downloaded data
    and the corpus database information are stored. If this directory
    does not exist, PyThaiNLP will automatically create a new one.
+   
    By default, it is specified to the directory called ``pythainlp-data``
    within the home directory.
+   
+   Type `thainlp data path` at command line to see current PYTHAINLP_DATA_DIR.

pythainlp/__main__.py

@@ -6,20 +6,19 @@
 
 
 def main(args=None):
-    """The main routine of PyThaiNLP command line."""
+    """ThaiNLP command line."""
     if args is None:
         args = sys.argv[1:]
 
     parser = argparse.ArgumentParser(
-        usage="pythainlp command [subcommand] [options]"
+        "thainlp", usage="thainlp <command> [options]"
     )
 
     parser.add_argument(
         "command",
         type=str,
-        default="",
-        nargs="?",
-        help="[%s]" % "|".join(cli.COMMANDS),
+        choices=cli.COMMANDS,
+        help="text processing action",
     )
 
     args = parser.parse_args(sys.argv[1:2])

pythainlp/cli/__init__.py

@@ -1,26 +1,22 @@
+"""Command line helpers."""
 import sys
+from argparse import ArgumentParser
 
-from . import corpus
-from . import tokenize
-from . import soundex
-from . import tag
+from . import data, soundex, tag, tokenize
 
-# commands should be verb when possible
-COMMANDS = sorted(['corpus', 'tokenize', 'soundex', 'tag'])
+# a command should be a verb when possible
+COMMANDS = sorted(["data", "soundex", "tag", "tokenize"])
 
 CLI_NAME = "thainlp"
 
 
-def make_usage(s):
-    prog = f"{CLI_NAME} {s}"
+def make_usage(command: str) -> dict:
+    prog = f"{CLI_NAME} {command}"
 
-    return dict(
-        prog=prog,
-        usage="%(prog)s command [subcommand] [options]"
-    )
+    return dict(prog=prog, usage=f"{prog} [options]")
 
 
-def exit_if_empty(d, parser):
-    if not d:
+def exit_if_empty(command: str, parser: ArgumentParser) -> None:
+    if not command:
         parser.print_help()
         sys.exit(0)

pythainlp/cli/corpus.py

@@ -0,0 +1,86 @@
+"""
+thainlp dataset/corpus management command line.
+"""
+import argparse
+
+from pythainlp import cli, corpus
+from pythainlp.tools import get_pythainlp_data_path
+
+
+class App:
+    def __init__(self, argv):
+        parser = argparse.ArgumentParser(
+            prog="data", usage="thainlp data <subcommand>",
+        )
+        parser.add_argument(
+            "subcommand",
+            type=str,
+            choices=["catalog", "info", "get", "rm", "path"],
+            help="action on dataset/corpus",
+        )
+        args = parser.parse_args(argv[2:3])
+        getattr(self, args.subcommand)(argv)
+
+    def get(self, argv):
+        parser = argparse.ArgumentParser(
+            description="Download a dataset",
+            usage="thainlp data get <dataset_name>",
+        )
+        parser.add_argument(
+            "dataset_name", type=str, help="dataset/corpus's name",
+        )
+        args = parser.parse_args(argv[3:])
+        if corpus.download(args.dataset_name):
+            print("Downloaded successfully.")
+        else:
+            print("Not found.")
+
+    def rm(self, argv):
+        parser = argparse.ArgumentParser(
+            description="Remove a dataset",
+            usage="thainlp data rm <dataset_name>",
+        )
+        parser.add_argument(
+            "dataset_name", type=str, help="dataset/corpus's name",
+        )
+        args = parser.parse_args(argv[3:])
+        if corpus.remove(args.dataset_name):
+            print("Removed successfully.")
+        else:
+            print("Not found.")
+
+    def info(self, argv):
+        parser = argparse.ArgumentParser(
+            description="Print information about a dataset",
+            usage="thainlp data info <dataset_name>",
+        )
+        parser.add_argument(
+            "dataset_name", type=str, help="dataset/corpus's name",
+        )
+        args = parser.parse_args(argv[3:])
+        info = corpus.get_corpus_db_detail(args.dataset_name)
+        if info:
+            print(info)
+        else:
+            print("Not found.")
+
+    def catalog(self, argv):
+        """Print dataset/corpus available for download."""
+        corpus_db = corpus.get_corpus_db(corpus.corpus_db_url())
+        corpus_db = corpus_db.json()
+        corpus_names = sorted(corpus_db.keys())
+        print("Dataset/corpus available for download:")
+        for name in corpus_names:
+            print(f"- {name} {corpus_db[name]['version']}", end="")
+            corpus_info = corpus.get_corpus_db_detail(name)
+            if corpus_info:
+                print(f"  (Local: {corpus_info['version']})")
+            else:
+                print()
+
+        print("\nUse subcommand 'get' to download dataset.")
+        print("Example: thainlp data get crfcut")
+
+    def path(self, argv):
+        """Print path for local dataset."""
+        print(get_pythainlp_data_path())

pythainlp/cli/data.py

@@ -1,44 +1,34 @@
+"""
+thainlp soundex command line.
+
+Take input text from command line.
+"""
 import argparse
 
 from pythainlp import cli
-from pythainlp.soundex import soundex
+from pythainlp.soundex import DEFAULT_SOUNDEX_ENGINE, soundex
 
 
 class App:
-
     def __init__(self, argv):
-        parser = argparse.ArgumentParser("soundex")
-
+        parser = argparse.ArgumentParser(
+            prog="soundex",
+            description="Convert a text to its sound-based index.",
+        )
         parser.add_argument(
-            "subcommand",
+            "-a",
+            "--algo",
+            dest="algorithm",
             type=str,
-            nargs="?",
-            help="[udom83|lk82|metasound]"
+            choices=["udom83", "lk82", "metasound"],
+            help="soundex algorithm",
+            default=DEFAULT_SOUNDEX_ENGINE,
         )
-
         parser.add_argument(
-            "-t",
-            "--text",
-            type=str,
-            help="input text",
+            "text", type=str, help="input text",
         )
 
-        args = parser.parse_args(argv[2:3])
-
-        cli.exit_if_empty(args.subcommand, parser)
-        subcommand = str.lower(args.subcommand)
-
-        cli.exit_if_empty(args.text, parser)
-
-        sdx = ""
-        if subcommand.startswith("udom"):
-            sdx = soundex(args.text, engine="udom83")
-        elif subcommand.startswith("lk"):
-            sdx = soundex(args.text, engine="lk82")
-        elif subcommand.startswith("meta"):
-            sdx = soundex(args.text, engine="metasound")
-        else:
-            raise NotImplementedError(
-                f"Subcommand not available: {subcommand}")
+        args = parser.parse_args(argv[2:])
 
+        sdx = soundex(args.text, engine=args.algorithm)
         print(sdx)